Biosynthesis of organic photosensitizer Zn-porphyrin by diphtheria toxin repressor (DtxR)-mediated global upregulation of engineered heme biosynthesis pathway in Corynebacterium glutamicum

Zn-porphyrin is a promising organic photosensitizer in various fields including solar cells, interface and biomedical research, but the biosynthesis study has been limited, probably due to the difficulty of understanding complex biosynthesis pathways. In this study, we developed a Corynebacterium glutamicum platform strain for the biosynthesis of Zn-coproporphyrin III (Zn-CP III), in which the heme biosynthesis pathway was efficiently upregulated. The pathway was activated and reinforced by strong promoter-induced expression of hemAM (encoding mutated glutamyl-tRNA reductase) and hemL (encoding glutamate-1-semialdehyde aminotransferase) genes. This engineered strain produced 33.54 ± 3.44 mg/l of Zn-CP III, while the control strain produced none. For efficient global regulation of the complex pathway, the dtxR gene encoding the transcriptional regulator diphtheria toxin repressor (DtxR) was first overexpressed in C. glutamicum with hemAM and hemL genes, and its combinatorial expression was improved by using effective genetic tools. This engineered strain biosynthesized 68.31 ± 2.15 mg/l of Zn-CP III. Finally, fed-batch fermentation allowed for the production of 132.09 mg/l of Zn-CP III. This titer represents the highest in bacterial production of Zn-CP III reported to date, to our knowledge. This study demonstrates that engineered C. glutamicum can be a robust biotechnological model for the production of photosensitizer Zn-porphyrin.

. A schematic overview describing the target pathway and overall metabolic strategies for Znporphyrin production in C. glutamicum. (a) The heme biosynthesis pathway used in this study. Glucose indicates a major sole energy source used in this study for cell growth and porphyrin production. Glycolysis starting from glucose is a chain of reactions to form pyruvate, which enters tricarboxylic acid (TCA) cycle with acetyl-CoA. TCA cycle is a series of biochemical reactions providing essential biochemical intermediates, such as 2-oxoglutarate, succinyl-CoA, etc., for cellular energy production and amino acid biosynthesis containing l-glutamate and l-glycine. Zinc-porphyrin biosynthesis is initiated from l-glutamate. The red letter represents the gene overexpression in C. glutamicum. The T-shape and red X-shape denote the gene repression by the transcriptional regulator and inactivation of the HrrA protein by gene repression by the DtxR protein, respectively. The solid-line arrow and dashed-line arrow indicate the general metabolic pathway and abbreviated metabolic pathway, respectively. The green arrow denotes enhanced metabolic flux by the combinatorial overexpression of hemA M , hemL and dtxR genes. The width of each green arrow is proportional to the result of the mRNA expression level of heme biosynthesis genes when the DtxR protein was overexpressed with HemA M and HemL protein. Purple and black colored chemicals are the target product Zn-CP III and porphyrin chassis CP III, respectively. (b) The regulatory mechanism of the transcriptional regulator DtxR in C. glutamicum. The DtxR protein and HrrA protein bind in the promoter region in front of target genes, respectively, repress the gene expression of those 43,45,47 . Major symbols were described in the dashed black line box. The black line arrow shows processing of regulatory mechanisms of the HrrA protein. Abbreviations: GSA, l-glutamate 1-semialdehyde; ALA, 5-aminolevulinic acid; PBG, porphobilinogen; HMB, hydroxymethylbilane; UPG III, uroporphyrinogen III; CPG III, coproporphyrinogen III; CP III, coproporphyrin III; Zn-CP III, zinc-coproporphyrin III; UP III, uroporphyrin III; HemA, glutamyl-tRNA reductase; HemL, glutamate-1semialdehyde 2,1-aminomutase; HemB, porphobilinogen synthase; HemC, hydroxymethylbilane synthase; HemD, uroporphyrinogen-III synthase; HemE, uroporphyrinogen decarboxylase; HemY, coproporphyrinogen III oxidase; DtxR, diphtheria toxin repressor; HrrA, heme responsive regulator activator.
SCiEntifiC REPORtS | (2018) 8:14460 | DOI: 10.1038/s41598-018-32854-9 ALA production, corresponding to 600.79 ± 62.47 mg/l of ALA after 72 h of cultivation ( Supplementary Fig. S2). Furthermore, while the control strain showed no production of any porphyrins except heme, the CGPB02 strain produced 33.32 ± 12.26 mg/l of uroporphyrin III (UP III), 23.68 ± 4.98 mg/l of CP III, and 16.56 ± 0.98 mg/l of heme, which was 7.79-fold, 3.04-fold, and 1.61-fold more than the CGPB01 strain, respectively (Fig. 2). Additionally, 22.38 ± 1.37 and 73.56 ± 17.89 mg/l of total porphyrins were produced in CGPB01 and CGPB02, respectively. Maximum cell growth was not affected by the expression of hemA M and hemL genes under the tac promoter ( Supplementary Fig. S2). This result indicates that the expression of hemA M and hemL genes is important for activation of the heme biosynthesis pathway in C. glutamicum, and the expression under the tac promoter was strong enough to increase the metabolic flux of the heme biosynthesis pathway compared with the trc-derived promoter.
Enhanced metabolic flux toward a porphyrin chassis CP III by the combinatorial expression of the hemA M , hemL and dtxR genes. In this study, we selected CP III as a final chassis for Zn-porphyrin biosynthesis. It has been revealed that not only CP III acted as a chelate of zinc, but also Zn-CP III was biosynthesized from CP III produced in bacteria including wild type Paracoccus denitrificans and Streptomyces sp. in culture medium supplemented with zinc sulfate 24,40 . As mentioned above, the transcription regulation mechanism between partial heme operon and the HrrA protein in C. glutamicum has been studied, although the cell growth of the hrrA gene deletion strain was significantly decreased 43,45 . Hence, we hypothesized that the DtxR protein may serve as an effective enhancer of the heme biosynthesis pathway. To verify this hypothesis, the dtxR gene (derived from C. glutamicum) encoding the DtxR protein was introduced into the control strain and the CGPB02 strain, named the CGPB03 and CGPB04 strains, respectively. Contrary to expectations, while the production of CP III and UP III in the CGPB03 strain was not detected, the CGPB04 strain showed a 1.82-fold and 2.17-fold increase (60.90 ± 3.26 and 51.35 ± 1.69 mg/l) of UP III and CP III production over the CGPB02 strain, respectively (Fig. 3a, Supplementary Fig. S3). Although the overexpression of the hemA M and hemL genes somewhat delayed the exponential growth phase, whether or not the dtxR gene was present, the maximum cell growth and glucose consumption were not affected (Fig. 3b,c). These results show that the combinatorial expression of dtxR with hemA M and hemL not only overcomes the rate-limiting step but also improves CP III production by enhancing the metabolic flux of the heme biosynthesis pathway without the critical negative effect of cell growth.
To enhance the metabolic flux of the heme biosynthesis pathway toward CP III, the hemA M , hemL and dtxR genes were additionally cloned in an E. coli-C. glutamicum shuttle vector pEKEx2 containing the tac promoter and pBL1 origin of replication. A pBL1 origin of replication contained in Shuttle vector pEKEx2 shows 8-30 copy numbers per cell, and a pCG1 origin of replication contained in pMT-tac shows 13-22 copy numbers per cell 48,49 . Expression of the hemA M , hemL and dtxR genes using the pEKEx2 vector was more effective at producing CP III than using the pMT-tac vector. Compared with the CGPB04 strain, the engineered strain harboring pEKEx2::he-mA M LdtxR (named the CGPB07 strain herein) showed a 1.22-fold increase in CP III production, corresponding to 62.54 ± 2.09 mg/l (Fig. 3a). Interestingly, UP III production in the CGPB07 strain was decreased 1.55-fold compared to the CGPB04 strain ( Supplementary Fig. S3). The result means that the metabolic flux from UP III to CP III may be extended by the reinforced expression of hemA M , hemL and dtxR genes through the pEKEx2 vector. Therefore, the CGPB07 strain is more useful for Zn-CP III production compared to other engineered strains.
The DtxR protein uses two ferrous iron as an essential cofactor 45 . Hence, cultivation with various concentrations of iron was performed on the CGPB04 and CGPB07 strains for enhanced production of CP III. However, along with increasing the iron concentration, CP III production in both the CGPB04 and CGPB07 strains decreased constantly while heme production was increased (Fig. 3d, Supplementary Fig. S4). These results mean that ferrous iron added in the culture medium may be used in processes to biosynthesize heme from CP III.

Global upregulation of the heme biosynthesis pathway by the transcriptional regulator
DtxR. It has been reported that the mRNA expression level of hemA, hemC (encoding porphobilinogen deaminase), hemE (encoding uroporphyrinogen decarboxylase), hemY (encoding protoporphyrinogen oxidase) and hemL was decreased when the hrrA gene was deleted in C. glutamicum 43 . However, transcriptional analysis of hemB and hemD (encoding uroporphyrinogen-III synthase) genes in the hrrA gene deletion strain as well as the transcriptional regulation of the heme biosynthesis pathway by the overexpression of the DtxR protein has not been studied yet. In particular, the DtxR protein has never been used for porphyrins production, to our knowledge. To investigate the positive effect of the DtxR protein in enhancing the metabolic flux of the heme biosynthesis pathway, we analyzed the relative mRNA expression of heme biosynthesis-related genes when the DtxR protein was overexpressed. Compared to the CGPB02 strain, the mRNA expression levels of the hemA, hemL, hemC, hemD, hemE and hemY genes in the genome directly involved in CP III production after the precursor synthesis were all increased in the CGPB04 strain, which showed 10.83 ± 1.21-, 1.77 ± 0.46-, 1.48 ± 0.22-, 4.50 ± 0.55-, 2.15 ± 0.27-and 3.23 ± 1.19-fold increases respectively (Fig. 4). Furthermore, the mRNA expression level of the hrrA gene in the CGPB04 strain was significantly decreased 0.28 ± 0.02-fold over the CGPB02 strain. However, the hemB mRNA expression level did not change, and this result means that the hemB gene may not be regulated by the HrrA protein. Taken together, these data indicate that genes related to the heme biosynthesis pathway were mostly upregulated by hrrA gene repression via overexpression of the DtxR protein, which may result in increasing the total porphyrin production in the engineered strain. In addition, the mRNA expression levels of the hemA, hemL, hemC, hemD, hemE and hemY genes in the CGPB07 strain were more increased than those of the CGPB04 strain, and showed the 35.55 ± 3.37-, 1.89 ± 0.18-, 37.40 ± 8.58-, 5.79 ± 0.41-, 22.02 ± 2.20-and The cultivation of all engineered strains was conducted in triplicate, and all data and error bars denote mean values of three independent experiments and standard deviations of mean values, respectively. Symbols: square, CGPB02 strain; circle, CGPB03 strain; triangle, CGPB04 strain; inverted triangle, CGPB07 strain; orange color, CP III production; green color, heme production; yellow color, cell growth; white color, glucose.
SCiEntifiC REPORtS | (2018) 8:14460 | DOI:10.1038/s41598-018-32854-9 7.92 ± 1.88-fold increase over the CGPB02 strain, respectively. The hrrA mRNA expression level in the CGPB07 strain was more decreased than that of the CGPB04 strain, showed the 0.14 ± 0.01-fold decrease over the CGPB02 strain. The result means that the mRNA expression of heme biosynthesis-related genes may be improved by the hrrA gene repression by the improved expression of the DtxR protein through the pEKEx2 vector.
Zn-CP III biosynthesis by engineered C. glutamicum strains by the combinatorial expression of the hemA M , hemL and dtxR genes. For Zn-CP III production, CP III-producing engineered strains were shake-cultivated in a glucose-minimal medium supplemented with zinc sulfate for 72 h, and most engineered strains except for the CGPB03 strain successfully produced Zn-CP III (Fig. 5a). Compared with the CGPB02 strain that produced 33.54 ± 3.44 mg/l of Zn-CP III, the CGPB04 and CGPB07 strains showed 1.65-and 2.04-fold increases in Zn-CP III production, corresponding to 55.44 ± 1.58 and 68.31 ± 2.15 mg/l, respectively. Similarly, as in the case of CP III, the CGBP07 strain showed the highest titer of Zn-CP III in bacteria reported to date, and the growth pattern and glucose consumption of all engineered strains were rarely affected (Fig. 5b,c).
Interestingly, relative to the total titer of CP III produced in the CGPB07 strain cultured without zinc sulfate, the total titer of Zn-CP III produced in the strain cultured with zinc sulfate was increased to 68.31 ± 2.15 mg/l from 62.54 ± 2.09 mg/l of CP III without certain genetic modifications, corresponding to a further 4.4% increase (Figs 3a, 5a). On the other hand, the total titer of heme produced in the strain cultured with zinc sulfate was decreased compared to that produced in the strain cultured without zinc sulfate, from 17.08 ± 0.10 mg/l to 2.64 ± 0.32 mg/l (Fig. 5d). This result means that CP III was partially used in Zn-CP III production rather than in heme production. The absorption spectrum of metabolites produced in the CGPB07 strain when zinc sulfate was supplemented showed not only the strongest feature at 406 nm but also minor features at 538 and 574 nm, representing optical characteristics of Zn-CP III ( Supplementary Fig. S5) similar to that of previous studies 24,40 . Typically, CP III has a maximum Soret band at 393 nm, consistent with the absorption maximum of metabolites produced in the strain cultured without zinc sulfate. These results suggest that engineered C. glutamicum strains successfully produced Zn-CP III, and efficiently used robust porphyrin metabolism to biosynthesize Zn-CP III.

Fed-batch fermentation for enhanced biosynthesis of Zn-CP III in engineered C. glutamicum.
To test the potential of the CGPB07 strain for large-scale production of Zn-CP III, we carried out 1.8-l fed-batch fermentation in a 5 l fermenter (Fig. 6). Glucose feeding was initiated after 15 h of fermentation and was additionally injected whenever the glucose concentration was below 20 g/l. During fed-batch fermentation of the CGPB07 strain, the maximum titer reached 132.09 mg/l of Zn-CP III in 82 h of fermentation in which the total input of glucose was 350.5 g. The overall yield and productivity were 1.27 mg/g glucose and 1.61 mg/l/h, respectively. The specific growth rate in the exponential growth phase (between 8 and 20 h) was 0.12/h. The maximum OD 600 was 131.33, corresponding to the measured cell biomass of 37.04 g/l at the end of fermentation.

Discussion
Much research has recently advanced Zn-porphyrin and porphyrin derivatives as promising raw material for solar cell devices, solid/liquid interface and photodynamic therapy over the past several years [1][2][3][4][5][6] . Until now, the development of porphyrin production in bacteria has lagged behind precursor ALA production presumably because of the complex synthetic processes and regulatory mechanisms of the heme biosynthetic pathway. Toriya et al. found that wild type Streptomyces sp. strain AC8007 biosynthesizes Zn-CP III (10 mg/l), but there was no metabolic design for overproduction of Zn-CP III in this wild type strain 40 . Kwon et al. reported successful production of CP III, PP IX and heme (89 ± 9 μmol/l of total porphyrins, corresponding to approximately 50 mg/l of total porphyrins) in E. coli by overexpressing eight genes related to heme biosynthesis 37 . However, the introduction of many genes and vectors may require excessive use of antibiotic resistance markers and reduce cell growth by metabolic burden 50 . Indeed, engineered E. coli strains, named as the ECPB01, ECPB02 and ECPB03 strains, less produced Zn-CP III than engineered C. glutamicum strains, corresponding to 1.93 ± 0.05 mg/l, 1.87 ± 0.07 mg/l and 1.84 ± 0.05 mg/l, respectively ( Supplementary Fig. S6). The porphyrin production in the engineered E. coli strain overexpressing the HemA M and HemL proteins (C5 pathway) in this study was similar to that in the engineered strain overexpressing the ALA synthase (C4 pathway) used in the study of Kwon et al. 37 . Furthermore, the DtxR protein only worked in C. glutamicum, but not in E. coli (Fig. 3a, Supplementary Figs S3, S6). This result means that C. glutamicum is suitable for Zn-CP III and have the possibility of performing the further metabolic engineering. In the field of microbial engineering, pertinent gene expression and regulation of the target biosynthesis pathway are important factors for high-yield production of a target product 51 . In this study, we confirmed the potential of C. glutamicum as a platform strain for Zn-CP III and porphyrin production. The activation and upregulation of the long and complex heme biosynthesis pathway by the transcriptional regulator DtxR alongside HemA M and HemL proteins in C. glutamicum can allow for efficient biosynthesis of Zn-CP III. Furthermore, target gene expression by the efficient genetic tool such as replacement of the strong promoter and origin of replication achieved the high production of CP III, Zn-CP III and Heme. Furthermore, the mRNA expression of most heme biosynthesis genes except for the hemB gene was significantly improved in the CGPB07 strain (Fig. 4). Interestingly, while the hemD gene, which is located at 500-bp upstream of the hemB gene, was upregulated by the hrrA gene repression, the hemB gene expression was not affected in both the CGPB04 and CGPB07 strains. Even though the hemB gene is close to the hemD gene in genome, their expression may be regulated by different expression factors containing a promoter and repressor because it is predicted that they do not exist as an operon (an operon was predicted at http://www.coryneregnet.de/) 43 . Based on these results, It can be speculated that hemB and hemD genes are individually expressed by different factors and that there exists no binding site for the HrrA protein in the promoter region of the hemB gene. Frunzke et al. has identified that the HrrA protein bound in the promoter region of hemA and hemE genes although the promoter region of hemD and hemB genes remained unclear 43 . Nonetheless, the enhanced mRNA expression level of most heme biosynthesis genes by the combinatorial expression of hemA M , hemL and dtxR genes induced substantial porphyrin production (Figs 3a, 4, Supplementary Fig. S3). Interestingly, in the CGPB03 strain overexpressing the dtxR gene alone, most of the porphyrins except for heme were not produced ( Supplementary Fig. S3). We suggest the reason may be because the native HemA protein overexpressed by the DtxR protein is feedback-inhibited by the final product heme [41][42][43] .
In terms of the insertion of a zinc ion into metal-free porphyrin, CP III may be more suitable for Zn-porphyrin synthesis than PP IX because CP III has the characteristic of spontaneous chelation of a zinc ion without any energy requirement while PP IX requires an enzyme reaction for Zn-PP IX biosynthesis 24,52,53 . Indeed, as described in our results regarding CP III, Zn-CP III and heme concentration, CP III generated for subsequent biosynthesis pathways was used for Zn-CP III production instead of heme production (Figs 3, 5). Dailey et al. recently revealed that the Actinobacteria and Firmicutes uses CP III instead of PP IX in heme biosynthesis 54 . C. glutamicum, which belongs to phylum Actinobacteria, is able to biosynthesize heme from a coproporphyrin-dependent pathway. As reported in Fig. 3, CP III accumulation was shown in metabolites produced in the CGPB02, CGPB04, and CGPB07 strains. However, none of the engineered strains constructed in this study produced PP IX despite increased mRNA expression levels of the hemN gene (encoding coproporphyrinogen III oxidase) ( Supplementary  Fig. S7). Furthermore, we confirmed that PP IX was produced when coproporphyrinogen III oxidase derived from E. coli with hemA M , hemL and dtxR genes was heterogeneously expressed in C. glutamicum ( Supplementary  Fig. S8). This observation suggests that C. glutamicum may not have a protoporphyrin-dependent pathway for heme biosynthesis, and therefore is suitable in terms of being a coproporphyrin-producing platform strain.
Even though the mRNA expression levels of the hemA, hemL, hemB, hemC, hemD, hemE and hemY genes for CP III production, as well as that of the hemH gene, were increased, heme production in the CGPB04 strain was slightly decreased, compared with the CGPB02 strain (Supplementary Figs S3, S7). Regarding this result, we speculate that the DtxR protein and heme competed for ferrous ion as a cofactor and required metal ions, respectively. Although the heme production from CP III was increased in the culture supplemented with iron sulfate of the CGPB07 strain, the bioconversion yield from CP III to heme was still low at 31.55 ± 3.39% (Fig. 3d). It is tempting to speculate that the final biosynthetic section from CP III to heme may be a major bottleneck for heme biosynthesis. Kwon et al. suggested that the limited transport of a metal ion into the cell and the low activity of the HemH protein in vivo may limit heme biosynthesis in the recombinant E. coli 37 . It has been recently reported that heme biosynthesis in the coproporphyrin-dependent pathway is performed by the HemQ protein which decarboxylates coproheme resulting in heme [54][55][56][57] . Hence, a mechanistic and enzymatic perspective of the coproporphyrin-dependent pathway in C. glutamicum is required for enhanced heme production. Nevertheless, our result regarding heme concentration in the CGPB07 strain showed the high-level production, corresponding to 29.57 ± 2.46 mg/l (Fig. 3d).
In summary, we demonstrated the possibility of C. glutamicum as a platform strain for the microbial production of Zn-porphyrin Zn-CP III as well as heme, and the titer of the porphyrins in the final engineered strain CGPB07 are the highest, reported to date in bacteria (Figs 3d, 6, Supplementary Fig. S3). These bio-based Figure 6. The fed-batch fermentation profile of the CGPB07 strain for Zn-CP III production. The x-axis, left y-axis, right y-axis, and right y-offset denote time, OD 600 , glucose, and Zn-CP III produced, respectively. Symbols: inverted triangle, CGPB07 strain; purple color, Zn-CP III production; yellow color, cell growth; white color, glucose. compounds can be used as a valuable material in various research related to solar cell photosensitizers, biological sensors as well as biomedical application such as antimicrobial, antitumor and anticancer photodynamic therapy [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]58,59 . In further study, large scale production of Zn-porphyrin in bacteria by microbial and metabolic engineering can be directly applied to solar cell technology and biomedical research. Indeed, Toriya et al. has directly applied Zn-CP III biosynthesized from wild type Streptomyces sp. used in their previous study to photodynamic therapy for antitumor effect in mice 40,59 . To enhance porphyrin biosynthesis and increase the diversity of types of porphyrins, further microbial engineering in C. glutamicum should be performed, including rational pathway engineering of the endogenous metabolism toward increased availability of an additional rate-limiting step, fine-tuning of enzyme engineering for improving target enzymes representing low activity, and the development of transport systems for efficient substrate supply.

Materials and Methods
Reagents and chemicals. Cell culture media used in this study was purchased from Daejung chemical Co.

Strains and plasmids.
All plasmids and bacterial strains constructed and used in this study are listed in Table 1. E. coli DH5α was chosen as the cloning host strain, and C. glutamicum ATCC 13826 was used in the construction of recombinant strains and experiments for porphyrin production in this study. S. typhimurium, E. coli and C. glutamicum strains were used to obtain hemA, hemL and dtxR genes from genomic DNA, respectively.
All oligonucleotides for the construction of recombinant vectors were listed in Supplementary Table S1. pMT-tac and pEKEx2 vectors were used as E. coli and C. glutamicum shuttle vectors and for the expression of target genes in C. glutamicum. To express hemA M and hemL genes under the control of the tac promoter, the hemA M and hemL gene operon was amplified with HemA M BamHI F and HemL NotI R primers using pMT-trc::he-mA M L vector as a template. The hemA M -hemL gene fragment was ligated into BamHI-NotI restriction sites of the pMT-tac vector by T4 DNA ligation, resulting in the pMT-tac::hemA M L vector. For the construction of pMT-tac::hemA M LdtxR and pMT-tac::dtxR vectors, the dtxR gene was amplified with DtxR NotI F and DtxR NotI R primers using pMT-tac::hemA M L and pMT-tac vectors as a template, respectively. To prevent self-ligation by one-cut restriction digestion, the digested sample was treated with Cloned Alkaline Phosphatase (Takara, Japan). The dtxR gene fragment was ligated into the NotI restriction sites of pMT-tac::hemA M L and pMT-tac vectors, resulting in pMT-tac::hemA M LdtxR and pMT-tac::dtxR vectors. In the same manner, the pEKEx2::hemAL vector was constructed by means of restriction digestion and ligation into PstI-SalI sites of the pEKEx2 vector. The dtxR gene was also introduced into SalI-BamHI sites of pEKEx2 and pEKEx2::hemA M L vectors, resulting in pEKEx2::dtxR and pEKEx2::hemA M LdtxR. Except for an adjacent gene from a promoter, a ribosomal binding site (AAGGAG) was inserted between downstream of the stop codon and upstream of the start codon in individual genes. Transformation of all recombinant vectors in E. coli and C. glutamicum was performed by the methods previously described 27 . Culture medium and conditions. For the construction of engineered strains, E. coli and C. glutamicum were grown overnight in shaking cultivation at 37 °C and 30 °C, respectively, and 200 rpm in Luria-Bertani medium. C. glutamicum wild type and engineered strains were grown overnight in shaking cultivation at 30 °C and 150 rpm in brain-heart infusion (BHI) -. The seed culture for porphyrin production was washed once with modified CGXII medium and transferred to main flask cultivation at a starting OD 600 of 1. Main flask cultivation for porphyrin production was conducted in 500 ml baffled flasks containing 100 ml modified CGXII medium containing 40 g/l glucose and 1 mM Isopropyl β-d-1-thiogalactopyranoside (IPTG). For glutamate overproduction, 6 U/ml penicillin G was supplemented after 12 h of cultivation. Modified CGXII medium consists of 20 g (NH 4 ) 2 SO 4 , 5 g urea, 1 g KH 2 PO 4 , 1 g K 2 HPO 4 , 42 g 3-morpholinopropanesulfonic acid (MOPS), 0. For fed-batch fermentation, stock cells were streaked onto BHI agar plates with 25 mg/l kanamycin and cultured for 30 h of standing incubation at 30 °C. One colony was used to inoculate 20 ml of the first seed culture and grown overnight in shaking cultivation at 30 °C and 150 rpm. The first seed culture was transferred to 200 ml of baffled flasks containing the second seed culture. The second seed culture was washed once with modified CGXII medium and transferred to fermenter containing 1.8 l modified CGXII medium, 40 g/l glucose, 1 mM IPTG and 25 mg/l kanamycin. The fermentation was performed at 30 °C and 300 to 600 rpm. The pH was automatically maintained at 6.8 by 4 M NaOH and 5 M H 2 SO 4 . Air aeration rate and dissolved oxygen were maintained at 1 vvm and 30% of air saturation, respectively. The foam was removed by the addition of 1:10 diluted antifoam Y-30 emulsion (Sigma-Aldrich). When glucose was lower than 20 g/l, 50% glucose stock solution was manually fed.
Analytical methods. Cell growth of all strains used in this study was measured at OD 600 using a UV-vis spectrophotometer (Mecasys Co., Ltd., Korea). The dry cell weight was experimentally determined on the basis of the correlation model OD 600 1 = 0.282 g dry cell weight (DCW)/l and used for the calculation of biomass concentration of fed-batch fermentation. Glucose concentration was estimated using a Glucose (GO) Assay Kit (Sigma-Aldrich, USA). For measurement of intracellular metabolites, 1 ml of cell pellet was harvested by centrifugation (13,000 rpm at 4 °C for 3 minutes), and the supernatant excluding the cell pellet was used for the measurement of extracellular metabolites. For the extraction of the metabolite from cells, including inter alia, heme and other porphyrins, the cell pellet was disrupted using-modified actone:HCl extraction methods described by Espinas et al. 60 . After 1 ml of acetone:HCl (95:5) buffer was added to the cell-harvested tube, the mixture was vortexed and diluted with 1 ml of 1 M NaOH. The intracellular sample which was disrupted and supernatant was filtered using an MCE (Mixed Cellulose Ester) filter (Hyundai micro, Korea) for concentration analysis. Porphyrin concentration was determined using a high-performance liquid chromatography (HPLC) system (Waters Corporation, USA), which consists of a dual λ absorbance detector (Waters 2487), binary HPLC pump (Waters 1525), and autosampler (Waters 2707). The filtered sample was separated in a SUPELCOSIL ™ LC-18-DB HPLC Column 5 μm particle size, L × I.D. 250 × 4.6 mm (Supelco Inc., USA) using a linear gradient method of 20-95% solvent A in B at 40 °C. Solvent A is a 10:90 (v/v) HPLC grade methanol:acetonitrile mixture, and solvent B is a 0.5% (v/v) trifluoroacetic acid (TFA) in HPLC grade water. The flow rate was 1 ml/min for 40 minutes, and the absorbance was determined at 400 nm. The data were estimated using Waters Empower-3 software. ALA concentration was measured using the colorimetric assay called Ehrlich's reagent 61 . One volume of the supernatant was chemically reacted with 0.5 volume of 1 M sodium acetate buffer (pH 4.8) and 0.25 volume of acetylacetone at 100 °C for 15 minutes. After that, the reagent mixture was cooled in an ice bath for 15 minutes. The modified Ehrlich's reagent was added to the cooled mixture and stayed at room temperature for 30 minutes. The absorbance was determined at 554 nm using Epoch 2 microplate spectrophotometer (BioTek Instruments, USA). RNA Preparation and qRT-PCR. To prepare total RNA, the CGPB02 and CGPB04 strains were cultivated in modified CGXII medium and harvested at 12 h. Total RNA in the cell pellet was extracted using a Hybrid-R prep kit (GeneAll, Korea) according to the manufacturer's instructions. Total RNA with random primer was denatured at 65 °C for 10 minutes, and reverse-transcriptionally synthesized to cDNA templates by M-MLV reverse transcriptase. All oligonucleotides for qRT-PCR were designed based on target gene sequence (Supplementary  Table S2). qRT-PCR analysis was determined using a StepOnePlus thermocycler (Thermo Fisher Scientific, USA) with HiPi Real-Time PCR 2x Master Mix (SYBR green) according to the manufacturer's instructions. The mRNA level of each gene of the CGPB04 strain was calculated relative to those of the CGPB02 strain using the comparative ΔΔCt method and was always normalized using the 16S rRNA level as an internal standard 62 .